LLMs don't learn.
They may exhibit features of intelligence, but to the degree that they do, it is because they are an interface to the (static!) body of human knowledge they were trained on. They reflect, and are entirely bound by, the static body of human knowledge that forms their training base.
